Cost Modeling and Estimation for OLAP-XML Federations

The ever-changing data requirements of today’s dynamic businesses are not handled well by current OnLine Analytical Processing (OLAP) systems. Physical integration of unexpected data into OLAP systems is a long and time-consuming process, making logical integration, or federation, the better choice in many cases. The increasing use of XML, e.g. in business-to-business (B2B) applications, suggests that the required data will often be available in XML format. Thus, federations of OLAP and XML databases will be very attractive in many situations. However, a naive implementation of OLAP-XML federations will not perform well enough to be useful. In an efficient implementation, cost-based optimization is a must, creating a need for an effective cost model for OLAP-XML federations. However, existing cost models do not support such systems. In this paper we present a cost model for OLAP-XML federations, along with techniques for estimating the cost model parameters in a federated OLAP-XML environment. The paper also present the cost models for the OLAP and XML components in the federation on which the federation cost model is built. The cost model has been used as the basis for effective cost-based query optimization in OLAP-XML federations. Experiments show that the cost model is precise enough to make a substantial difference in the query optimization process.